This research addressed an urban traffic intelligent control system,which adopts a multi-agents coordination in urban traffic control to coordinate the signal of adjacent intersections for eliminating the congestion of traffic network.An agent represents a signal intersection control,and multi-agents realize coordination of multiple intersections to eliminate congestion.Based on Recursive Modeling Method and Bayesian learning that enables an agent to select his rational action by examining with other agents by modeling their decision making in conjunction with dynamic belief update.Based on this method,a simplified multi-agent traffic control system is established and the results demonstrate its effectiveness.It is very important for ITS.